Yushu Wang Xingxing: The biggest mistake made in the past ten years is not believing enough in AI

According to Kuai Technology News on October 12, today, the 26th China International Industrial Expo was held in Shanghai. At the opening ceremony, Wang Xingxing, chairman, general manager and chief technology officer of Yushu Technology, said that with the rapid improvement of AI model capabilities, the technical critical point of embodied intelligence may come faster than expected, and it may even occur within one or two years if it is faster. Especially after the release of GPT-6 some time ago, the awareness of practitioners in the entire embodied intelligence industry has also been greatly improved. "We put it on various robots for various task tests, and the results are actually very good." Wang Xingxing revealed that he personally feels that the entire critical point will arrive relatively quickly, if not within one or two years, because GPT-7 will be released next year, which may directly solve the problem of embodied intelligence. According to him, Yushu Technology and other robot companies are currently conducting pilot projects in factories. Currently, large-scale applications are needed, and both generalization and efficiency must be further improved. He believes that although the current large-scale application of robots in factories still faces problems such as generalization capabilities and efficiency, they are expected to be solved in the next few years. "The biggest mistake I made in the past decade or so was that I didn't believe enough in AI. The implementation of commercial services may be faster than in the industrial field." Wang Xingxing said that his confidence in AI is increasing year by year. In addition, Wang Xingxing also said that many domestic and foreign robot companies have already carried out pilot projects in factories, but they still need to solve the problems of generalization capability and efficiency before large-scale application. He believes that these problems are expected to be solved in the next few years. Judging from the application progress of different scenarios, the application progress of commercial service scenarios may be faster than that of industrial scenarios, while the advancement of home scenarios may be relatively late.